Just rewrote the fore-mentioned patch using TCG ops. Here's some numbers
running the tests on
my local machine:
- using current master:
(1/1) tests/avocado/boot_linux_console.py:BootLinuxConsole.test_ppc_powernv8:
PASS (71.00 s)
(1/1) tests/avocado/boot_linux_console.py:BootLinuxConsole.test_ppc_powernv8:
PASS (69.57 s)
(1/1) tests/avocado/boot_linux_console.py:BootLinuxConsole.test_ppc_powernv8:
PASS (76.04 s)
(1/1) tests/avocado/boot_linux_console.py:BootLinuxConsole.test_ppc_powernv9:
PASS (72.62 s)
(1/1) tests/avocado/boot_linux_console.py:BootLinuxConsole.test_ppc_powernv9:
PASS (76.50 s)
(1/1) tests/avocado/boot_linux_console.py:BootLinuxConsole.test_ppc_powernv9:
PASS (73.58 s)
- after my TCG Ops rewrite to count instructions:
(1/1) tests/avocado/boot_linux_console.py:BootLinuxConsole.test_ppc_powernv8:
PASS (39.97 s)
(1/1) tests/avocado/boot_linux_console.py:BootLinuxConsole.test_ppc_powernv8:
PASS (40.19 s)
(1/1) tests/avocado/boot_linux_console.py:BootLinuxConsole.test_ppc_powernv8:
PASS (41.76 s)
(1/1) tests/avocado/boot_linux_console.py:BootLinuxConsole.test_ppc_powernv9:
PASS (40.88 s)
(1/1) tests/avocado/boot_linux_console.py:BootLinuxConsole.test_ppc_powernv9:
PASS (41.49 s)
(1/1) tests/avocado/boot_linux_console.py:BootLinuxConsole.test_ppc_powernv9:
PASS (42.04 s)
Also, there's a high possibility that the code I wrote is not optimized since
I'm not well
versed with TCG ops/code. I expect that after a couple of reviews from Richard
we might be able
to bring down those numbers even further.
This is behaving like 6.2. We should be fine (until we add more counters :)
I'll clean this up and send for review.
ok. We might have a last ppc PR in 2021.
Thanks a lot,
C.
